One argument that you might be able to make is that home prices follow urban density. New York City, for example, is dense. And homes in New York City tend to be more expensive than those in, oh I don’t know, rural Canada. So with this, you might conclude that development and density are bad — it makes housing more expensive. But then there’s places like San Jose, California. It’s not very dense, and yet it has some of if not the most expensive housing in the US.

Well, it turns out that housing density and median housing values don’t actually exhibit a particularly strong correlation. A better and much stronger relationship can be found in what Kasey Klimes explains, here, in this excellent post, which is that home prices more accurately follow incomes. In other words, the more high paying jobs that exist in a market, the more likely that housing will be expensive.

Here is what that looks like for US metros over 1 million people:

The above chart compares median home value to aggregate income per unit of housing. And here, Kasey discovers an r-value of 0.9, which suggests that “over 81% of median home values in large metros can be attributed to aggregate income per unit of housing.” This explains why San Jose, and San Francisco, are such outliers. They have very high incomes for every unit of available housing, despite the former being not all that dense.

Okay, so now that we know this, how do we make housing more affordable? One option is to just make people poorer. If you reduce incomes per unit of housing, then home prices will, almost certainly, go down. And this is why poorer cities tend to have more affordable housing. But this is obviously suboptimal. The better option is to keep people wealthy and simply increase the denominator in “aggregate income per unit of housing.”

Meaning: build more housing!

Chart: Kasey Klimes